Anticoagulation in chronic kidney disease: from guidelines to clinical practice

Abstract Background Chronic kidney disease (CKD) is a major global public health problem, being closely connected to cardiovascular disease. CKD involves an elevated thromboembolic risk and requires anticoagulation, but the high rates of hemorrhage render it quite challenging. Hypothesis There are no consensus recommendations regarding anticoagulation in CKD. Due to the currently limited data, clinicians need practical clues for monitoring and optimizing the treatment. Methods Based on the available data, this review outlines the benefit‐risk ratio of all types of anticoagulants in each stage of CKD and provides practical recommendations for accurate dosage adjustment, reversal of antithrombotic effect, and monitoring of renal function on a regular basis. Results Evidence from randomized controlled trials supports the efficient and safe use of warfarin and direct oral anticoagulants (DOACs) in mild and moderate CKD. On the contrary, the data are poor and controversial for advanced stages. DOACs are preferred in CKD stages 1 to 3. In patients with stage 4 CKD, the choice of warfarin vs DOACs will take into consideration the pharmacokinetics of the drugs and patient characteristics. Warfarin remains the first‐line treatment in end‐stage renal disease, although in this case the decision to use or not to use anticoagulation is strictly individualized. Anticoagulation with heparins is safe in nondialysis‐dependent CKD, but remains a challenge in the hemodialysis patients. Conclusions Although there is a need for cardiorenal consensus regarding anticoagulation in CKD, adequate selection of the anticoagulant type and careful monitoring are some extremely useful indications for overcoming management challenges.

at 69.8% vs 34.8% in the general population. 1 Also, if microalbuminuria is detected and glomerular filtration rate (eGFR) is less than <60 mL/min/1.73m 2 , there is an increased risk of cardiovascular events and mortality. The relationship is early and gradual across the entire spectrum of CKD (Table 1), causing the majority of patients to succumb to cardiovascular complications. 2 The risk of atrial fibrillation (AF) and acute coronary syndrome (ACS) is double in patients with eGFR <60 mL/min/1.73m 2 . 3 The study model most relevant to clinical practice is AF in CKD. The prevalence of AF increases with the decline in renal function, the prevalence of end-stage renal disease (ESRD) being two or even three times higher than in the general population. 4 Depending on the studied cohorts, the prevalence of AF in CKD patients was estimated to range between 12% and 18% compared to 7%-8% in the general population over 65 years of age. The prevalence of AF remains high (11.6%) in dialysis-dependent CKD, and 12 months after kidney transplantation, the risk of AF occurrence increases up to 35.6% per 1000 patient years. 5 On the other hand, large prospective cohort studies demonstrate the relationship between incident AF and an 80% higher risk of decline in eGFR and a 116% higher risk of proteinuria detection. 6 In the chronic renal insufficiency cohort prospective study, AF led to a 3-fold higher risk of progression to ESRD. 7 The association between AF and CKD engenders a much higher thromboembolic risk, which in case of ischemic stroke ranges from 26% to 49%, depending on the study. 5 Also, the relative risk of mortality increases by up to 66%. 6 The association between ACS and CKD is also well documented in registries. Thus, 40% of non-ST segment-elevation myocardial infarction (NSTEMI) cases and 30% of ST segment-elevation myocardial infarction (STEMI) cases associate eGFR <60 mL/min/1.73 m 2 , and mortality is double compared with general population. 3 The risk of pulmonary venous thromboembolism (VTE) in CKD increases by 25%-30% is constant in all CKD stages, and typically characterizes the nephrotic syndrome. 8

| WHAT IS THE UNDERLYING CONNECTION BETWEEN CKD AND CVD?
The bidirectional relationship between CVD and CKD is due to the intervention of major cardiovascular risk factors shared by both disorders. Concurrently, CKD entails factors that are frequently involved in uremia, such as systemic inflammation, oxidative stress, activation of the renin-angiotensin-aldosterone system, malnourishment, anemia, microalbuminuria, hyperhomocysteinemia, as well as hyperparathyroidism, abnormalities in bone and mineral metabolism (in which phosphorus, fibroblastic/fibroblast growth factor 23, vitamin D deficiency play an important role), the particular profile of apolipoprotein isoforms, platelet hyperreactivity, all with impact on the cardiovascular system. 4,9,10 This results in a hypercoagulable state that generates arterial and venous thromboembolic complications, as well as the progression of kideny disease. The pathophysiological substrate involves the three components of Virchow's triad-stasis and turbulent blood flow, vascular endothelial injury, and hypercoagulability. Endothelial injury/dysfunction is an essential promoter of the proinflammatory, procoagulant and pro-proliferative state. 9,10 In case of uremia, high levels of fibrinogen, thrombin-antithrombin complexes, thrombomodulin, von Willebrand factor, plasminogen activator inhibitor 1 (PAI-1), factor VII are markers of endothelial dysfunction. A particular mechanism in uremia is the intervention of homocysteine via thrombin activation, fibrin formation, and reduced release of tissue plasminogen activator from the endothelium. Via these mechanisms, along with PAI-1, homocysteine causes a decreased fibrinolytic activity. 9 Also, the platelets of uremic patients are dysfunctional due to the activation of certain microRNA-altering mechanisms. This results in microparticles expressing tissue factor, the key-element for the initiation of the coagulation cascade. PAI-1 secretion links endothelial dysfunction to structural cardiovascular (CV) changes, as it promotes tissue fibrosis. Arterial stiffness occurs along with early onset atherosclerosis and vascular calcifications.
The heart shows left ventricular hypertrophy and marked myocardial fibrosis with impact on coronary circulation, atrial and ventricular remodeling, and blood flow implicitly. 4 The clinical consequence is an increased thromboembolic risk.
The paradox in CKD is the association between the high thromboembolic risk and major hemorrhagic risk with declining kidney function. Platelet hyperreactivity in the early stages is replaced by decreased platelet activity and impaired platelet-vessel wall interaction. This is caused by the alteration of platelet-dependent mechanisms involved in physiological hemostasis. 9,10 Overall, platelet adhesion and aggregation are reduced. 11 The prohemorrhagic state is potentiated by CKD-related anemia, extrinsic iatrogenic factors (nonsteroidal anti-inflammatory drugs, antithrombotic medication, antibiotics, invasive procedures, dialysis methods), and gastrointestinal lesions. 4 Epidemiologic studies confirm the elevated hemorrhagic risk, which can be 4.1 times higher in ischemic stroke and 10.7 times higher in intracerebral hemorrhage in dialysis patients. 12 In CKD, the fragile balance between the risk of thromboembolic events and hemorrhage puts the population requiring anticoagulation treatment in a very difficult position for the following reasons: • the need for anticoagulants is much higher in CKD; • the population with advanced CKD stages is frequently excluded from controlled randomized trials, so there is no consistent evidence for the efficacy and safety of anticoagulants; • there are no thromboembolic and hemorrhagic risk scores that adequately define individual risk; • the risk-benefit ratio is influenced by numerous variables specific to this subgroup; • there are pharmacokinetic and pharmacodynamic features related to the impaired renal functions, and interaction with other drugs requiring adjustment of therapeutic regimens; • the methods used to assess renal dysfunction vary between the studies, which obviously leads to contradictory results; • there is no consensus on the recommendations for oral anticoagulation, and the use of a particular type of anticoagulant, especially in stages 4 to 5 CKD cannot be supported.   17 Dose adjustment is also necessary because in the liver plasma half-life (t 1/2 ) is shortened, the renal clearance is enhanced, and the interaction with other drugs is changed. Warfarin administration is even more difficult in dialysis patients, being associated with an increased risk of thromboembolic events and hemorrhage. 13,16 Ultimately, a particular problem is the relationship between warfarin-vascular calcifications-renal function decline. The mechanism entails vitamin K inhibition that indirectly inhibits matrix G1a protein, thus promoting vascular calcification and calciphylaxis.
Progression of renal vascular calcifications is associated with renal function decline and higher hemorrhagic and thromboembolic risk. 9,16 Direct oral anticoagulants (DOACs) are a therapeutic option with evident advantages. However, in case of renal dysfunction, their use makes dose adjustment mandatory, as there is a variable degree of renal clearance (Table 3). 13 Oral that recommend an additional reduction of rivaroxaban doses (10-mg once daily (QD)) and edoxaban (25-mg QD) in severe CKD. 9 Also, the FDA issued a black box warning against the use of edoxaban in patients with CrCl >95 mL/min, considering the numerical but not statistically significant excess of ischemic strokes. 9 Rivaroxaban is also not recommended in patients with VTE and CrCl <30 mL/min, and for CrCL 30 to 49 mL/min the recommended dose is 15-mg BID for 21 days followed by 20-mg QD. 22 For patients on hemodialysis, but not for stage 5 nondialysis patients, the FDA allows apixaban 5-mg BID, although according to pharmacokinetic data and label recommendations the dose of 2.5-mg BID would ensure adequate plasma concentration. 9,23,24 Labeling supports that rivaroxaban may be administered at a dose of 15 mg QD. 24 There is no conclusive data for edoxaban.
One particular adverse effect is DOACs-induced nephropathy, experimentally demonstrated and described by isolated reports; it is produced by the tubular obstruction caused by hematic cylinders, as well as by the activation of protease-activated receptor 1. 25  preferred AF rate control. 26 Another study conducted in the United States found that in spite of guidelines and FDA-issued recommendations, 60% of the patients with mild-to-moderate CKD receive lower doses of DOACs, which could account for the excess thromboembolic events. 27

| Heparin treatment
Heparin, no matter of type, is administered according to the classic rules, depending on the associated disorder (acute coronary syndrome/VTE). Dose adjustment is necessary in advanced CKD and is primarily based on guideline recommendations (Table 4). [28][29][30] Unfractionated heparin (UFH) is preferred because it has a short half-life that allows for the anticoagulant effect to wear off within 1 to 4 hours, even in patients with severe renal dysfunction at high hemorrhagic risk. In addition to this, there is an antidote (protamine) used to rapidly reverse the effects of UFH, although the guidelines recommend UFH in severe CKD without adjustment of impaired renal clearance and interpatient variability of accumulation. Therefore, nephrology practice recommends decreasing the initial standard dose by 33%, and subsequently dose adjustment based on aPTT. 16 In the NSTEMI guideline, indications are primarily based on dose adjustments. 29 The primary argument is the high hemorrhagic risk per se in severe CKD.
Low-molecular-weight heparins (LMWHs) are preferred owing to their pharmacokinetic predictability, ease of administration without the need for monitoring. Renal clearance is indirectly proportional to molecular weight, therefore it requires dose adjustments in CKD stages 4 and 5 (Table 4). For dosage adjustment purposes, it is recommended to monitor the activity of antifactor Xa (anti-Xa level) in order to avoid underdosage and achieve optimal therapeutic level, respectively. 16 Dosing indications are the result of either small-scale open-label studies, or analysis of CKD subgroups in the randomized trials, adopted by guidelines. Enoxaparin is the most commonly used low-molecular-weight heparin (LMWH) and the 1-mg/kg QD regimen recommended in severe CKD the most studied. There is no data for dalteparin and tinzaparin in severe CKD; therefore, it is preferable to avoid administering them. 16 Although preferred in cases of heparininduced thrombocytopenia, fondaparinux is not recommended in severe CKD.

| Monitoring the treatment with heparins
Activated Partial Thromboplastin Time (aPTT)-based dose adjustment is still recommended for UFH to achieve an optimal therapeutic level (aPTT range 1.5-2.0). 16  to a prevalence of 6.42 to 9.91/1000 patient years. 34 In predialysis patients, the rates range between 4% and 21%. 20 Renal dysfunction is even more frequent in acute coronary syndromes, being present in 30% of STEMI and 40% of NSTEMI patients. 20 For VTE, the risk is estimated to increase from 29% in mild CKD up to 134% in patients on dialysis. 8 Overall, CKD is an independent risk factor for ischemic and hemorrhagic stroke, estimated at around 5%-6%/year. 34 Compared to the general population, the thromboembolic risk is 2.5 to 5.5 times higher depending on renal function status, while the hemorrhagic risk is at least double. 4,9,34 The practical consequence is that although prophylactic anticoagulation is necessary, the benefits and safety may be affected compared to the general population.
Literature data on VKAs primarily refer to warfarin, which practically entails the extrapolation of data to other coumarin drugs. Most of the evidence comes from the analysis of AF patient subgroups.
There is only one randomized control study (Stroke Prevention in Atrial Fibrillation III study) that included patients with stage 3 of CKD (42%) and analyzes warfarin compared to the population with normal renal function. 9 Data analysis in CKD subgroup show that well-adjusted doses reduce the risk of ischemic stroke and systemic embolism by 76% and 67%, respectively, without statistically significant differences in major bleeding rates. Other data on warfarin derive from registries and observational studies that include CKD subgroups.
Overall, the results are consistent in terms of effectiveness in reducing the thromboembolic risk, the risk of cardiovascular/all-cause mortality, as well as the risk of a fatal stroke. 34  ESRD on dialysis. 35 Consistent across studies is the assertion of an increased risk of hemorrhage in dialysis patients. 4,9,34  DOACs serve as a viable alternative to VKAs. As mentioned before, the available evidence for daily practice results from the analysis of subgroups with mild-to-moderate CKD in landmark trials. 9,15 For severe CKD, clinical data are lacking and pharmacokinetic studies recommend dose reduction. 15 In ESRD, although rivaroxaban and apixaban appear to be safe, prospective clinical data are needed. 15 Also, the meta-analysis by Nielsen  Consequently, the initiation of anticoagulation therapy in severe CKD and ESRD is still a debate topic. Practitioners are confused by the information based on observational studies with irregular designs or inaccurate administrative registries, with extremely variable results. 9 Systematic analyses and meta-analyses can no longer provide accurate data, even when conducting sophisticated data analysis.
Uncertainty is even higher in VTE, for which there are no guideline recommendations. Another problem is that of OACs for primary prevention of thromboembolic events in AF, particularly in stage 5 and dialysis patients. 9,34 Cardiologists tend to extrapolate guideline recommendations to patients without CKD, while nephrologists are reticent in this respect.
The latest 2011 KDIGO (Kidney Disease: Improving Global Outcomes) guidelines state that only nephrologists should recommend OACs as primary prevention in ESRD and dialysis patients, based on a strictly individualized algorithm. A prudent approach to secondary prevention is to give OAC to all ESRD patients, without absolute contraindications. 40 The use of DOACs is nonstandardized, possible in low apixaban/rivaroxaban doses in patients with ESRD at very high risk of ischemic stroke and with warfarin intolerance or suboptimal TTR, or in patients with recurrent stroke on adequate warfarin treatment. 40 There are several ongoing trials, comparing DOACs and warfarin in reports, the data require validation. 15

| CONCLUSIONS
There are therapeutic resources for CKD requiring anticoagulation, but evidence from randomized controlled trials is limited for mild and moderate stages. DOACs should be avoided in severe forms of CKD.
Although VKAs have the downside of being unpredictable, the major advantage is that their use can be extended all the way through the terminal stage and can be easily reversed if necessary. When faced with such a complex patient, initiating and maintaining anticoagulation becomes a challenging task, which needs to be based on the primum non nocere principle. 34 The strict determination of the individual profile, adequate selection of the type of anticoagulant and careful monitoring are some extremely useful indications for overcoming management challenges. Although there is a need for cardiorenal consensus, based on the current data the fact remains that DOACs are preferred in stages 1 to 3. In stage 4, the choice between DOACs vs warfarin will consider the pharmacokinetics of the drugs and patient characteristics. Warfarin remains the first-line treatment in ESRD, although in this case the actual decision to use or not to use anticoagulation is strictly individualized. Anticoagulation with heparin is safe in nondialysis-dependent CKD if optimal monitoring is ensured, but remains a challenge in the hemodialysis patients.